Column
Suplementary material
A cohort study in an inner district with high incidence of tuberculosis in Barcelona
The source code is available at github/jrealgatius
by Jordi Real & Ray Puig
Sampling and follow-up sheeme
Sampling
The sampling method used was density sampling. This plot shown a time scheme of the recruitment of cases with an example of 5 pairs of cases. Incident Diabetes cases were selected as they occurred and their respective controls of the same sex, age on inclusion date.
A case is a person who is recorded as diabetes at index date. A control is a person who has not yet have diabetes at the index date of the case. The index date of the case can also be the onset of a disease (Diabetes), in which case controls are persons who do not yet have the disease (are not yet exposed).
Follow-up
Once identified, they are followed until the end of follow-up. Which determines your time at risk (Black line). The follow-up was until 12/31/2018, until death or until switch group
Switch group
Example of change of group
Flow-Chart
Flow-Chart
Flow chart of how the sample finally analyzed has been arrived at. Initially we started from a potential population from Barcelona of 85,919, of which 10,065 Diabetics (Prevalent as incidents) + 75,854 potential controls. Of these, 8004 diabetics and 8004 controls of the same age and sex were matched.
Incidence
Incidence rates
During the follow-up, we collected data on Tuberculosis event recorded.
Number of subjects, tuberculosis events, and incidence rate (per 100.000 persons-year) and confidence interval (95%CI), Overall, by groups and type of diabetes during the follow-up
| Group | Subjects (N) | Persons-years | Tuberculosis events | Incidence rate | (95% CI) |
|---|---|---|---|---|---|
| Overall | 16008 | 129804 | 73 | 56.2 | (44.1;70.7) |
| Control | 8004 | 61198 | 25 | 40.9 | (26.4;60.3) |
| Diabetes | 8004 | 68605 | 48 | 70.0 | (51.6;92.8) |
| DM1 | 355 | 3505 | 2 | 57.1 | (6.9;206.1) |
| DM2 | 7649 | 65101 | 46 | 70.7 | (51.7;94.3) |
| DM Incident | 3641 | 22965 | 23 | 100.2 | (63.5;150.3) |
| DM Prevalent | 4363 | 45640 | 25 | 54.8 | (35.4;80.9) |
Cumulative incidence curve by group
Sensitivity analysis
Sensitivity analysis
Additional analyses in order to check the robustness of the results include the different approach models
| Model | Adjust variables |
|---|---|
| 1 | Origin |
| 2 | Origin, Visits |
| 3 | Origin, CKD, BMI, Visits |
| 4 | Origin, Alcholism, Smoke, CKD, BMI, Visits |
Methods
Competing risk: Competing Risks Regression for Clustered Data. Regression modeling of subdistribution hazards for clustered right censored data. Failure times within the same cluster are dependent. crrs R package Vesion 1.1. This method extends Fine-Gray proportional hazards model for subdistribution (1999) to accommodate situations where the failure times within a cluster might be correlated since the study subjects from the same cluster share common factors This model directly assesses the effect of covariates on the subdistribution of a particular type of failure in a competing risks setting.
Bingqing Zhou and Aurelien Latouche (2013). crrSC: Competing risks regression for Stratified and Clustered data. R package version 1.1. https://CRAN.R-project.org/package=crrSC
Cox PH by clusters: Fits a Cox proportional hazards regression model with clusters. coxph function from {survival} R packages Therneau T (2015). A Package for Survival Analysis in S. version 2.38, <URL: https://CRAN.R-project.org/package=survival>.
Full model
Full model
